结合小波变换和独立分量分析的脑电特征提取

在多道脑电信号的独立分量分析 ICA (Independent Com ponent Analysis)过程中 ,脑电独立源在数量上的不确定性往往是造成分离结果不理想的主要原因。目前还没有找到解决此问题的有效方法。根据脑电信号自身的特点 ,提出了一种结合小波变换 WT(Wavelet Transform)和独立分量分析的脑电特征提取新方法— WICA。新方法的主要思路是先对每一道脑电数据进行小波分解 ,然后根据需要选择不同尺度的小波子带信号进行脑电独立分量提取。实验结果表明 ,WICA方法在一定程度上能够克服因脑电独立源的不确定性所带来的影响 ,并能获得较好的脑电特征提取效果。

EEG Pattern Extraction of Combining Wavelet Transform and Independent Component Analysis

In the process of independent component analysis (ICA) of multi-channel EEG signals, the uncertainty of the number and location of EEG independent sources will bring some difficulties to ICA application in EEG analysis. Until now, there is no effective method to solve this problem.According to the characteristics of EEG signal, WICA, a new approach of combining wavelet transform and ICA, is proposed. The main idea of this new method is that each EEG channel is first decomposed into several sub-band signals using wavelet transform, ICA algorithm is then applied to the sub-band signals at different wavelet scale to get the independent components of EEG. The experiment results show that WICA can overcome to some extent the difficulties caused by the uncertainty of EEG source and get better results of EEG pattern recognition.